The Ferranti Atlas 1 was a very significant development in the history of computing. Chiefly famous for being the machine which introduced the now, almost universal technique of paging to the world, it was, briefly, claimed as the world's fastest computer. It was said at the time that, when the Atlas 1 failed, the UK’s computing capacity was halved. Many other innovative hardware and software ideas first saw the light of day in Atlas 1, from instruction overlapping and spooling to the Compiler Compiler.
This emulator seeks to demonstrate the Atlas 1 order code by allowing you run Atlas 1 “jobs” one at a time. An Atlas 1 job would typically be to compile a program into store and execute it. In the emulator, in order to allow the user to visualise the process of execution, not to mention to support debugging techniques of which contemporary users could only have dreamt, execution takes place in an environment within what would appear to be a Windows-style debugger. No attempt, however, is made to emulate the whole machine which is beyond the capacity of the resources currently available. For example, there is no emulation of multiprogramming, although the one-level store concept can be seen in operation. The emulator is supplied with a number of Atlas 1 jobs some of which are for demonstration purposes and some of which generate components of the emulator itself.
At the time of writing the emulator is incomplete but very much demonstrable. Find out how much is (in)complete incomplete.html. I have obtained paper copies of the ABL library routines, of the Fixed Store and of the Brooker-Morris Compiler Compiler. However, I would welcome any information on some of the more obscure floating point function codes and any original ABL or CC programs would be useful.
The Computer Conservation Society is aware of the existence of a magnetic tape which may contain much of the system software of the Manchester Atlas 1. However, although we currently have no means of recovering the information from the tape we do have an Atlas 1 magnetic tape deck, albeit not currently working.
I would like to acknowledge the support I have received from Professor Simon Lavington and the late Dr Brian Napper both of whom were key members of the original Atlas team. Iain MacCallum and Bill Purvis have both made significant contributions to the rescue of the Brooker-Morris Compiler Compiler. I am grateful all of them. Recognition also to Brian Hardisty who wrote the ABL manual, a document far too good for its purpose, but truly excellent for mine. Finally, due credit should be given to Bob Hopgood, Chilton’s “Mr. Algol” (among many other things) who maintains the chilton-computing website (see below) and who has been unfailingly supportive.
Dik Leatherdale 2004-16
|Guide to Documentation||
The major thrust of the documentation available on this website/help file is a tutorial introduction to the emulator. Written largely for an audience which has some knowledge of Atlas 1, it nevertheless gives some detail of Atlas 1 and its Supervisor (operating system) where it is useful in explaining the operation of the emulator, particularly where the emulator by its very nature diverges from the real Atlas 1 (the emulator does not, of course, have any actual peripherals).
There is also a large amount of reference material, much of it in tabular summary form such as might be found in programming manuals.
Finally, a series of links to original material held in external websites which, it is hoped, may prove useful to readers.
|Where to go next||